Based on customhouse statistics of twenty-nine partners,trade misinvoicing for China is estimated in this paper.It is verified in events analysis that the indication of the estimations tracks the economic reform of Ch...Based on customhouse statistics of twenty-nine partners,trade misinvoicing for China is estimated in this paper.It is verified in events analysis that the indication of the estimations tracks the economic reform of China.It is indicated that the main channel for capital fight from China is export under-invoicing and import over-invoicing in 28 bilateral trades other than Hong Kong,while the import under-inwuing in the bilateral trades to Hong Kong results in a considerable illegal inflow of international portfolio capital.It is implied by the analysis that existent supervision of foreign trade is the shortage of capital control,and a reinforced custom system is the compensation for the national capital control system.展开更多
目的基于特征筛选算法探索利用胸腔积液和血清中CEA、CA125、CA153和CA199四种肿瘤标志物的组合与筛选对鉴别良、恶性胸腔积液的诊断价值。方法收集北京朝阳医院和武汉某医院收治的胸腔积液患者共452例,其中恶性胸腔积液患者143例、良...目的基于特征筛选算法探索利用胸腔积液和血清中CEA、CA125、CA153和CA199四种肿瘤标志物的组合与筛选对鉴别良、恶性胸腔积液的诊断价值。方法收集北京朝阳医院和武汉某医院收治的胸腔积液患者共452例,其中恶性胸腔积液患者143例、良性胸腔积液患者309例;取胸腔积液及配对血清标本,用化学发光法检测CEA、CA125、CA153和CA199浓度,辅以患者性别、年龄和医院所在城市三项人口学变量,首先应用独立性检验进行变量初筛,而后应用带惩罚项的逻辑回归和基于逻辑回归的模拟退火算法和遗传算法进行标志物筛选,根据受试者工作特征曲线下面积(area under the curve,AUC)和DeLong检验进行模型诊断效果的评估和比较。结果特征筛选结果以及回归系数和SHAP(shapley additive explanations)值一致表明胸腔积液CA199、CA153联合血清CEA为最优肿瘤标志物组合;在测试数据集上,该指标组合达到最高诊断精度(AUC=0.923),显著高于最优单标志物模型(AUC=0.877,P<0.001)和全标志物模型(AUC=0.906,P=0.044),灵敏度和特异度分别达到0.811和0.939。结论多项肿瘤标志物的联合应用相较单一标志物能够显著提升模型诊断精度,且合理的标志物筛选策略对提升诊断精度和简化模型有进一步帮助;本文推荐联合胸腔积液中CA199和CA153以及血清中CEA来建立诊断模型,并提供了该模型的列线图和实用化的网页计算器,为辅助临床诊断提供便利。展开更多
First of all,the definition of the statistical capacity for a country was given in this paper.Based upon it,author discussed how to evaluate a country’s statistical capacity both from input side and output side. A po...First of all,the definition of the statistical capacity for a country was given in this paper.Based upon it,author discussed how to evaluate a country’s statistical capacity both from input side and output side. A possible way of evaluating a country’s statistical capacity qualitatively was given.Finally it discussed how to building up statistical capacity for a country.展开更多
文摘Based on customhouse statistics of twenty-nine partners,trade misinvoicing for China is estimated in this paper.It is verified in events analysis that the indication of the estimations tracks the economic reform of China.It is indicated that the main channel for capital fight from China is export under-invoicing and import over-invoicing in 28 bilateral trades other than Hong Kong,while the import under-inwuing in the bilateral trades to Hong Kong results in a considerable illegal inflow of international portfolio capital.It is implied by the analysis that existent supervision of foreign trade is the shortage of capital control,and a reinforced custom system is the compensation for the national capital control system.
文摘目的基于特征筛选算法探索利用胸腔积液和血清中CEA、CA125、CA153和CA199四种肿瘤标志物的组合与筛选对鉴别良、恶性胸腔积液的诊断价值。方法收集北京朝阳医院和武汉某医院收治的胸腔积液患者共452例,其中恶性胸腔积液患者143例、良性胸腔积液患者309例;取胸腔积液及配对血清标本,用化学发光法检测CEA、CA125、CA153和CA199浓度,辅以患者性别、年龄和医院所在城市三项人口学变量,首先应用独立性检验进行变量初筛,而后应用带惩罚项的逻辑回归和基于逻辑回归的模拟退火算法和遗传算法进行标志物筛选,根据受试者工作特征曲线下面积(area under the curve,AUC)和DeLong检验进行模型诊断效果的评估和比较。结果特征筛选结果以及回归系数和SHAP(shapley additive explanations)值一致表明胸腔积液CA199、CA153联合血清CEA为最优肿瘤标志物组合;在测试数据集上,该指标组合达到最高诊断精度(AUC=0.923),显著高于最优单标志物模型(AUC=0.877,P<0.001)和全标志物模型(AUC=0.906,P=0.044),灵敏度和特异度分别达到0.811和0.939。结论多项肿瘤标志物的联合应用相较单一标志物能够显著提升模型诊断精度,且合理的标志物筛选策略对提升诊断精度和简化模型有进一步帮助;本文推荐联合胸腔积液中CA199和CA153以及血清中CEA来建立诊断模型,并提供了该模型的列线图和实用化的网页计算器,为辅助临床诊断提供便利。
文摘First of all,the definition of the statistical capacity for a country was given in this paper.Based upon it,author discussed how to evaluate a country’s statistical capacity both from input side and output side. A possible way of evaluating a country’s statistical capacity qualitatively was given.Finally it discussed how to building up statistical capacity for a country.
